AI Engineer

SHIELD is a device-first fraud intelligence platform that helps digital businesses worldwide eliminate fake accounts and stop all fraudulent activity. Powered by SHIELD AI, we identify the root of fraud with the global standard for device identification (SHIELD Device ID) and actionable fraud intelligence, empowering businesses to stay ahead of new and unknown fraud threats.We are trusted by global unicorns like inDrive, Alibaba, Swiggy, Meesho, TrueMoney, and more. With offices in LA, London, Jakarta, Bengaluru, Beijing, and Singapore, we are rapidly achieving our mission - eliminating unfairness to enable trust for the world.ResponsibilitiesAs an AI Engineer, you will work closely with the team to build and enhance AI-powered systems that support proactive identification of fraudulent behavior across our clientele's platforms. This is an opportunity to be part of a high-impact team that combines data, AI, and engineering to protect ecosystems.Build and maintain AI-powered systems for proactive fraud detection, including LLM-based and agentic workflowsDesign and implement RAG pipelines that ground models in SHIELD's fraud intelligence and data signalsDevelop and iterate on prompts and evaluations to ensure reliable, measurable model performanceRapidly prototype and ship new AI features, moving quickly from idea to productionExplore and integrate new data signals and model capabilities to improve fraud identification accuracyConduct comprehensive testing to ensure system reliability, performance, and cost-efficiencyWrite clear documentation for systems, prompts, workflows, and research findingsCollaborate closely with engineers and analysts to achieve shared project goalsRequirementsBachelor's Degree in Computer Science or a related field (or equivalent practical experience)Strong proficiency in Python (Go/Golang is a plus)Hands-on experience building with LLM APIs such as OpenAI, Anthropic (Claude), or Google Gemini — including prompting, RAG, or agentic workflowsExperience building and shipping production software, with a bias for moving fastExperience working with relational databases (e.g., MySQL, PostgreSQL)Familiarity with version control systems (e.g., Git)It will be good to have:Experience with vector databases or embedding-based retrievalExperience designing structured outputs from LLMs (e.g., JSON or schema-constrained generation)Familiarity with token optimization and cost/latency tuningFamiliarity with evaluation frameworks or methods for LLM outputsExperience working with caching systems (e.g., Redis, Memcached)Prior experience in the fraud detection or risk domain